This paper investigates the classification of a new type of Pseudo-Random (PN) codes which are generated using noise-like chaotic signals. The classification is made using differ- ent discriminant analysis classifiers such as linear, diagonal linear, quadratic, diagonal quadratic, and mahalanobis discriminant analysis classifiers. These classifiers are compared with other classifiers such as neural networks, support vector machines, k-nearest neighbor, and maximum likelihood classifiers. Higher order statistical (HOS) moments and cumulants of the eighth order are used as features. Simulation results illustrates the dependence of the proposed classification method on the type of the used classifier, the type of chaotic map and the initial values used for generating the chaotic code. Two types of one dimensional chaotic maps are considered in this work, namely, the logistic map and the Bended up down map. The performance, considered as the probability of correct decision, is shown to differ according to the type of the used classifier and is dependent on the signal-to-noise ratio. The results show that the quadratic discriminant analysis classifier outperforms the other discriminant analysis classifiers. Also, the performance of two codes generated from one logistic map with two different initial values outperforms the performance of two codes generated from one Bended up down map with two different initial values, two codes generated from logistic map with initial value and Bended up down map with the same initial value, and two codes generated from logistic map with one initial value and Bended up down map with another initial value.